vit_fold_3_v3
This model is a fine-tuned version of google/vit-base-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 1.1010
- Accuracy: 0.9647
- F1 Score: 0.9660
- Recall: 0.9685
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 150
- num_epochs: 100
- label_smoothing_factor: 0.15
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 Score | Recall |
|---|---|---|---|---|---|---|
| 2.6854 | 1.0 | 20 | 2.6942 | 0.3462 | 0.3547 | 0.3478 |
| 2.4323 | 2.0 | 40 | 2.5258 | 0.4519 | 0.4425 | 0.4417 |
| 2.1096 | 3.0 | 60 | 2.2691 | 0.5769 | 0.5725 | 0.5725 |
| 1.8003 | 4.0 | 80 | 1.9493 | 0.7308 | 0.7337 | 0.7288 |
| 1.5149 | 5.0 | 100 | 1.6536 | 0.8045 | 0.8100 | 0.8110 |
| 1.3394 | 6.0 | 120 | 1.4559 | 0.8622 | 0.8670 | 0.8727 |
| 1.2386 | 7.0 | 140 | 1.3304 | 0.8942 | 0.8957 | 0.8996 |
| 1.1850 | 8.0 | 160 | 1.2781 | 0.9135 | 0.9144 | 0.9176 |
| 1.1177 | 9.0 | 180 | 1.2214 | 0.9263 | 0.9283 | 0.9313 |
| 1.0805 | 10.0 | 200 | 1.1939 | 0.9359 | 0.9370 | 0.9386 |
| 1.0630 | 11.0 | 220 | 1.1824 | 0.9423 | 0.9437 | 0.9461 |
| 1.0597 | 12.0 | 240 | 1.1674 | 0.9423 | 0.9432 | 0.9448 |
| 1.0277 | 13.0 | 260 | 1.1526 | 0.9551 | 0.9563 | 0.9579 |
| 1.0241 | 14.0 | 280 | 1.1404 | 0.9551 | 0.9555 | 0.9546 |
| 1.0154 | 15.0 | 300 | 1.1305 | 0.9583 | 0.9592 | 0.9596 |
| 1.0142 | 16.0 | 320 | 1.1313 | 0.9487 | 0.9505 | 0.9542 |
| 1.0062 | 17.0 | 340 | 1.1250 | 0.9583 | 0.9599 | 0.9623 |
| 0.9936 | 18.0 | 360 | 1.1225 | 0.9583 | 0.9593 | 0.9584 |
| 0.9816 | 19.0 | 380 | 1.1186 | 0.9551 | 0.9557 | 0.9553 |
| 1.0086 | 20.0 | 400 | 1.1110 | 0.9583 | 0.9596 | 0.9597 |
| 0.9837 | 21.0 | 420 | 1.1251 | 0.9519 | 0.9533 | 0.9567 |
| 0.9834 | 22.0 | 440 | 1.1054 | 0.9551 | 0.9561 | 0.9566 |
| 0.9692 | 23.0 | 460 | 1.1031 | 0.9615 | 0.9620 | 0.9628 |
| 0.9973 | 24.0 | 480 | 1.1013 | 0.9551 | 0.9561 | 0.9566 |
| 0.9689 | 25.0 | 500 | 1.0996 | 0.9583 | 0.9592 | 0.9603 |
| 0.9655 | 26.0 | 520 | 1.0984 | 0.9551 | 0.9561 | 0.9566 |
| 0.9619 | 27.0 | 540 | 1.1071 | 0.9583 | 0.9598 | 0.9610 |
| 0.9746 | 28.0 | 560 | 1.1076 | 0.9519 | 0.9532 | 0.9554 |
| 0.9667 | 29.0 | 580 | 1.1114 | 0.9551 | 0.9562 | 0.9579 |
| 0.9629 | 30.0 | 600 | 1.1132 | 0.9487 | 0.9502 | 0.9530 |
| 0.9680 | 31.0 | 620 | 1.1108 | 0.9551 | 0.9565 | 0.9604 |
| 0.9747 | 32.0 | 640 | 1.1019 | 0.9551 | 0.9561 | 0.9566 |
| 0.9577 | 33.0 | 660 | 1.1087 | 0.9583 | 0.9593 | 0.9616 |
| 0.9670 | 34.0 | 680 | 1.1024 | 0.9551 | 0.9561 | 0.9566 |
| 0.9740 | 35.0 | 700 | 1.0980 | 0.9551 | 0.9562 | 0.9579 |
| 0.9668 | 36.0 | 720 | 1.0987 | 0.9615 | 0.9629 | 0.9647 |
| 0.9746 | 37.0 | 740 | 1.0923 | 0.9647 | 0.9659 | 0.9672 |
| 0.9585 | 38.0 | 760 | 1.1002 | 0.9615 | 0.9629 | 0.9647 |
| 0.9794 | 39.0 | 780 | 1.1101 | 0.9615 | 0.9630 | 0.9660 |
| 0.9628 | 40.0 | 800 | 1.0931 | 0.9647 | 0.9659 | 0.9672 |
| 0.9693 | 41.0 | 820 | 1.1010 | 0.9647 | 0.9660 | 0.9685 |
| 0.9635 | 42.0 | 840 | 1.1025 | 0.9615 | 0.9629 | 0.9647 |
| 0.9680 | 43.0 | 860 | 1.1042 | 0.9615 | 0.9629 | 0.9647 |
| 0.9670 | 44.0 | 880 | 1.0979 | 0.9615 | 0.9628 | 0.9634 |
| 0.9575 | 45.0 | 900 | 1.1008 | 0.9583 | 0.9592 | 0.9603 |
| 0.9705 | 46.0 | 920 | 1.1026 | 0.9583 | 0.9593 | 0.9616 |
| 0.9632 | 47.0 | 940 | 1.1008 | 0.9583 | 0.9593 | 0.9616 |
Framework versions
- Transformers 5.0.0
- Pytorch 2.10.0+cu128
- Datasets 5.0.0
- Tokenizers 0.22.2